ML Interference Prediction for Telecommunication Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Telecommunication networks face challenges in predicting and managing periodic interference, especially in scenarios with predictable traffic patterns and time division duplex operations, where existing methods struggle to accurately estimate interference from neighboring cells, leading to suboptimal radio resource management and link adaptation.
Innovation Solution
A method and apparatus that utilize machine learning models to predict future interference measurements by configuring communication devices to measure and analyze historical interference data, allowing network nodes to adjust configurations based on predicted interference values, thereby improving radio resource management and link adaptation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machine learning models are used to predict future interference measurements, then interference prediction accuracy is improved, but device complexity increases
Solution Approach 1:
The communication device performs preliminary actions by collecting historical interference measurements and training the machine learning model in advance. The model is trained using historical data stored in memory, enabling the device to predict future interference conditions before they occur, thus improving prediction accuracy while managing complexity through proactive preparation
Solution Approach 2:
The communication device serves itself by autonomously building and maintaining the machine learning model using its own historical interference measurements. The device independently trains the model, generates predictions, and provides these predictions to the network node without requiring external intervention, thereby managing complexity through self-sufficiency
2Measurement precision
If historical interference measurements are collected and analyzed, then prediction accuracy is improved, but loss of time increases
Solution Approach 1:
The device performs preliminary data collection and model training during periods when predictions are not immediately needed. Historical interference measurements are accumulated and the machine learning model is trained in advance, so that when predictions are required, the model is already prepared and can provide rapid predictions without causing time loss
Solution Approach 2:
The collection of historical interference measurements and model training is performed continuously over time rather than intermittently. This continuous accumulation of data and progressive model improvement ensures that the prediction accuracy enhances over time without requiring large batches of data to be processed at once, thereby reducing time loss
Data Source
AI summary
A method performed by a network node for a telecommunications network for handling interference variations for a communication device is provided. The method includes configuring the communication device to measure on a set of resources and to build a machine learning, ML, model to predict a future interference measurement from the set of resources. The method further includes signaling a request to the communication device to provide a prediction of the future interference measurement on the set of resources. The method further includes receiving, from the communication device, at least one prediction of future interference on the set of resources. The method further includes changing a network configuration for the communication device based on the received at least one prediction of future interference.